Year of Graduation
Non-Gaussian Models of Dependence in Asset Returns
Statistical Modelling and Actuarial Science
The study area of this research is the Non-Gaussian distributions of dependence in asset returns. There are empirical properties of asset returns which present a set of stylized empirical facts. These facts are shown that asset returns data mostly has Non-Gaussian distribution. Three models with non-Gaussian marginal are investigated on a daily return data for the 10 biggest Russian companies. The first model is Full rank Gaussian copula, the second model is a linear mixture of independent Levy processes. The third is correlates Gaussian components in a variance gamma representation. Also the concept of local correlation was introduced to detect the differences between models.